Real Temp - New temp program for Intel Core processors

[unclewebb]

oggie: Intel finally told users last year that the TJ Target for the E8x00 processors is 100C and TJMax is likely somewhere around there. I changed RealTemp to reflect that information.

The fact is that TJMax is not a fixed number. It varies from one E8600 to the next and even from one core to the next on the same CPU. That's very obvious on the 45nm Quads. Intel didn't release enough information about how much TJMax can vary so their release of TJ Target for some CPUs isn't much better than a number out of a hat. Find a number you like and check your calibration using rge's method to see if your choice or RealTemp's choice of TJMax is reasonable.

http://www.xtremesystems.org/forums/...postcount=2429

PRJ: RealTemp has the ability to shut down your computer based on an Nvidia GPU temperature. It won't do it automatically though. You would need to have it set it up more or less like this:



You would have to tell it in the Settings window to run RTShutdown.bat which runs the Windows shutdown.exe command for you. It's easy enough to edit that .bat file if you want to use some different parameters for that command or you can create your own .bat file and get RealTemp to do whatever you like. I thought I'd include one .bat file for an example.

You can turn off this feature and turn on the GPU log file and set the interval to 1 second and then go do some gaming. Once you're finished you'll be able to go back and see what are typical GPU temperatures for your setup. I liked these new features so much I went out and bought an Asus 9800GTX+ card recently to enjoy them. ATI makes it more of a hassle to read GPU temperatures so getting a different card was the easiest thing to do.

[PCTwin]

3.00 works fine for my old B3 qx6700. 3.06 only showed 2 cores and 3.07 only shows three cores.

[unclewebb]

3.06 only showed 2 cores and 3.07 only shows three cores.

Sounds like I'm making progress.

I intend to go back to the original code and original C++ compiler I was using for 3.00. There seemed to be a lot less of these random issues when using that. Hopefully I can add in the QX Load meter fix to 3.00 and the GPU button without screwing up anything else.

If 3.07 works on your computer then use it but if not then stick with 3.00. I might have to re-name it RealTemp darkzone because he seems to be the only one happy with it.

[darkzone]

I might have to re-name it RealTemp darkzone because he seems to be the only one happy with it.

Can you compile a personalized version for me? With my nick in title bar. Pliiizzzz [/joke]

[unclewebb]

Can you compile a personalized version for me?

I think I did already. It's called RealTemp 3.07. Enjoy.

Programming is fun. Every time you fix one bug you end up creating two new ones.

[darkzone]

I think I've just found a new motto for you: complicating the uncomplicated

[sabe]

3.07 works fine for me too by the way QX6700 and Q9550

[ChrisZ]

3.07 works fine for me too by the way QX6700 and Q9550

Works OK on QX9650 (Win XP).

[unclewebb]

http://www.fileden.com/files/2008/3/...507/RTCore.zip

Just doing some initial beta testing of the updated RealTemp / RivaTuner plug-in.



I added the ability to view individual load data for each core/thread to keep burebista happy.

Even with the polling interval dropped down to 0.2 seconds or 5 times a second, the average data is steady with two threads of Prime Small FFTs running while the individual loads are rapidly jumping up and down just like Intel intended. When you set the polling interval back to 1 second, it should look similar to the Task Manager graphs. More accurate, but similar.

I should have a fresh fix tomorrow for the 3 core Quad bug that I created when fixing darkzone's issues.

I haven't tested this plug-in with a hyper-threaded Core i7 yet so give it a try. You should have more data than you know what to do with it.

Unzip the files and copy them to your RealTemp directory and then click on the RivaTuner button in the RealTemp Settings window and install them by clicking on OK a couple of times.

[MrHydes]

uncleweb how about i7920 tjmax set?

[unclewebb]

how about i7920 tjmax set?

Can you translate that question to English?

Most of the Core i7 920 CPUs that I've seen have TJMax set to 100C if that's what you mean.

I saw in the forums yesterday that there are some new Core i7 CPUs coming out with CPUID 0x106A5 and Core Temp was showing one of them as TJMax = 92C. TJMax is contained within a register of these newer processors so a program like RealTemp or Core Temp can read that value directly and I'm pretty sure that both programs are doing that right.

If that didn't answer your question then try again.

[MrHydes]

Can you translate that question to English?

Most of the Core i7 920 CPUs that I've seen have TJMax set to 100C if that's what you mean.

I saw in the forums yesterday that there are some new Core i7 CPUs coming out with CPUID 0x106A5 and Core Temp was showing one of them as TJMax = 92C. TJMax is contained within a register of these newer processors so a program like RealTemp or Core Temp can read that value directly and I'm pretty sure that both programs are doing that right.

If that didn't answer your question then try again.

thanks uncle you've understand what i've meant to say... sorry

my offul english... I'll give a try after with new RT to riva, keep up the

good work for the community

[unclewebb]

http://www.fileden.com/files/2008/3/...alTempBeta.zip

I went back to the original 3.00 code and the old C++ compiler and added a few recent bug fixes and the GPU button to it. It includes support for the original Core i7 ES processors as well as the new ones. The 3.0x code is going into retirement.

Will the new / old RealTemp pass the darkzone test? Time for some beta testing to find out.

[oggie]

Unclewebb, love your work You must be proud to have made such a great application for all of us to use
Quick one if you dont mind, this 3.07 version....it works for GPU as well ?? You you please explain

Ignore above, have read through your thread and understand now

[unclewebb]

oggie: Unfortunately, RealTemp only supports Nvidia GPUs at the moment. Nvidia makes it simple to read temperature data but it's more of a hassle with ATI cards. I might add ATI support someday.

[oggie]

oggie: Unfortunately, RealTemp only supports Nvidia GPUs at the moment. Nvidia makes it simple to read temperature data but it's more of a hassle with ATI cards. I might add ATI support someday.

No probs Keep up your good work, all of us appreciate it

[DGill]

RealTemp 3.11 works well with i7. Thanks, much appreciated

[darkzone]

It seems that new version 3.11 doesn't work. Anyway, the big problem is my operating system which has an issue (I didn't found out what yet).

[PCTwin]

The missing core returns.

[maGSImum]

How accurate would these temperatures be considering the large distance to TjMax??



Ambient is 15 degrees Celsius btw., before you guys start saying that these themps are impossible...

[Hardass]

Running the 3.11 on a rig with Q6600 it shows load at 74% when all other programs show correct 100%??

[maGSImum]

Running the 3.11 on a rig with Q6600 it shows load at 74% when all other programs show correct 100%??

Same here with 3.00.. must have got something to do with realtemp controlling prime95.

Ow wait, you're not sensor testing... my bad.

Mine did just the same when sensor testing...

[Hardass]

Same here with 3.00.. must have got something to do with realtemp controlling prime95.

Mine is running F@H, but does same thing when running prime. Yes 3.00 did same.

[maGSImum]

Mine is running F@H, but does same thing when running prime. Yes 3.00 did same.

So if you're manually running prime95 it still displays 74%??

That's a problem I don't have.
It only occured to me that when sensortesting Realtemp is cutting the load and displays that correctly, but any other program is still displaying 100% (taskmanager etc.)

[burebista]

Mine looks fine.